1

CHAPTER ONE

INTRODUCTION

1.1 BACKGROUND

Successional management of large areas of natural vegetation depends on knowledge

of the composition of the vegetation, the extent to which it is being used, and the

changes that take place in response to differential use by herbivores and by fire

(Walker, 1976). Recently, more attention has been given to the Kalahari sand forests

of north western Matabeleland, which have so often been ignored. One of the major

management problems in the Kalahari sands woodlands is knowing whether

regeneration is adequate, given exploitation rates and effects of previous logging.

Knowledge of disturbance regimes is critical in understanding current and future

composition of forests (Gutierrez, Armesto and Aravena, 2004).

Numerous factors affect structure and composition of dry woodland. These factors

could be divided into biotic and abiotic. Abiotic factors include edaphic components

such as nitrogen and phosphorus concentrations (Stromgraad, 1992; Chidumayo,

1994), and disturbances such as natural fire (Kikula, 1986, Chidumayo, 1988a). Biotic

factors can be divided into natural occurring forces such as damage by herbivores and

anthropogenic factors including commercial charcoal production (Monela, O’kting

Ati and Kiwele, 1993), and collection of fuelwood (Abbot and Homewood, 1999).

Dynamic changes in dry savanna woodland systems are often interpreted to be the

result of disturbances, such as overgrazing, over browsing (Lubke and Thatcher,

1983; Pellew, 1983; Coughenour and Ellis, 1993), fire (Bell, 1982; Trollope, 1980,

2

Bond and Van Wilgen, 1996) and human utilisation (Blackmore, Mentis and Scholes,

1990; Shackleton, Griffin, Banks, Mavrandonis and Shackleton, 1994 (as cited by

Gotosa, 2002). Fire intensity influences regeneration of timber species by reducing

seedling establishment and recruitment of saplings to canopy (Gambiza, 2001).

Changes in population structure might indicate management impact before the loss of

a valuable timber species occurs. It may also alert managers to situations of declining

recruitment (Walker, Stone, Henderson and Vernede, 1986). Size class distributions

(Weiner and Caswell, 1977; Knowles and Grant, 1983) or stage class distributions

(Silvertown, 1987) are better indicators of reproductive output from a woodland or

forest.

Due to effect of fire and herbivory on seedling establishment, resprouting from stem

bases and roots are the major regenerative strategy in Baikiaeap plurijuga, commonly

called Zambezi Teak (Gambiza, 2001). Regeneration of B. plurijuga is poor (Piearce,

1986). Furthermore, its growth rate is low. This has led to speculations of future loss

of some forests due to lack of regeneration and recruitment. It has been hypothesized

that current extraction rates of exploitable stock would be exhausted at the turn of the

century (Mushove, 1991). Increasing demand for timber and timber products has led

to speculation that there is insufficient time for wood regeneration (Scholes and

Parsons, 1997).

While few studies have been carried out in Kalahari sand forests to document

ecological impacts of timber extraction, logging has been proved to have extensive

impacts on: floristic structure and composition, edaphic properties, and fauna. It is not

3

clear if this trend is discernible with B. plurijuga. The vegetation dynamics in

Kalahari sands are determined by, logging, fire, herbivory and interactions among

plants (Gambiza, 2001). Regeneration rate of B. plurijuga is slow, and disturbance

through felling, overgrazing and fire could lead to invasion by miombo species

(Bradley and Dewees, 1993).

1.2 Logging in the woodlands

Logging consists of periodic physical removal of portions of the dominant tree

canopy. This may benefit trees and shrubs that remain (Calvert, 1986a). In Zimbabwe,

the exploitation of commercial timber is carried out through private concessions, and

is regulated by the Forestry Commission under the Forest Act. There has been

concern over the decline of some indigenous timber species and their capacity to

regenerate after a disturbance. Increased demand for high quality timber has resulted

in extensive logging of important indigenous timber species. This has led to gene

impoverishment (Gondo, Nobanda and Mapaure, unpublished paper). Extraction of

wood for commercial sale and fuel is proceeding at an accelerated rate (Scholes and

Parsons, 1997). Indigenous timber trees that are harvested from woodlands are mainly

Zambezi teak and mukwa. These tree species maily occur in the Kalahari sands.

Logging has increased due to increased demand for timber for manufacturing. This is

creating pressure on these indigenous timber species.

1.3 Justification

With increased harvesting pressure on indigenous timber species due to increased

demand for timber and timber products in the Kalahari sands, it has become

imperative to study these tree species. Furthermore, the extent of logging needs to be

4

documented in light of increasing impacts of logging. Data on extent of logging could

help in decision-making pertaining to logging cycles. Many studies have concentrated

mainly on fire, yet logging is also a major form of disturbance in the indigenous

timber woodlands. Effects of prolonged logging are not fully understood. Data from

the present study will help in monitoring and assessment of impacts of logging on

regeneration of B. plurijug and species composition of the Gwayi Forest. For effective

management of Zambezi teak woodlands, there is need to monitor whether the rate of

timber extraction is sustainable. Such knowledge will help understand the dynamics

of species in the Gwayi Forest. This study seeks to assess regeneration, extent of

logging and changes in species composition in light of disturbances such as logging.

1.4 Objectives

This study aimed at developing baseline data on regenerative capacity of B

plurijuga under different logging regimes. The specific objectives were:

♦ to examine the size class structure of Baikiaea plurijuga in unlogged and logged

areas;

♦ to compare the extent of logging in two logged areas, one logged in 1995 and

another in 1999;

♦ to examine regeneration patterns in areas logged at different periods;

♦ to examine patterns of change in diversity and richness accompanying

regeneration;

♦ to describe species composition patterns, and their relationship with soil pH,

Nitrogen, Phosphorus and Potassium.

5

1.5 Research questions

♦ How do size class structure, diversity and richness compare in an unlogged area

and areas logged at different times (1995 and 1999)?

♦ How do extents of logging and pattern of regeneration compare in areas logged at

different times (1995 and 1999)?

♦ What is the relationship between species distribution and pH, Nitrogen,

Phosphorus and Potassium in areas logged at different times (1999 and 1995)?

1.6 Alternative hypotheses were:

H1: The size class structure of B. plurijuga is different in an unlogged area, area

logged in 1999 and area logged in 1995

H2: Regeneration patterns and diversity are different in unlogged area, area logged

in 1999 and area logged in 1995

6

CHAPTER TWO

LITERATURE REVIEW

2.1 Background

Baikiaea plurijuga is a deciduous tree that belongs to the family leguminosae

(Theilade, Sekeli, Hald and Graudal, 2001). It can grow to a maximum height of 20

metres and has a smooth bark, which becomes fissured and cracked later on. Leaves

are alternate, compound with four to five pairs of leaflets (Theilade, et al., 2001). Its

flowers are large and arranged in axial racemes up to 30 cm long. The bark is dark

brown and has dark brown buds. B. plurijuga starts flowering in December extending

into March with the peak in the middle of the rain season (Childes and Walker, 1987).

2.2 Zambezi Teak woodlands: definition, distribution and characteristics

Zambezi Teak woodlands are in fact dry savanna woodlands rather than true forests.

The vegetation in these woodlands are limited by available soil moisture within

rooting depth, seasonality of water availability, low nutrients and fire (Timberlake,

Nobanda and Mapaure, 1993).

The present distribution of Kalahari sands is azonal, and they are found from about

13˚S to 20˚ S, at least 6˚of the latitude (Huckabay, 1984). Kalahari sands cover 1 988

400 hactares world wide (Hogberg, 1986). Zambezi teak woodlands are found on

Kalahari sands. In Africa, they are found in Northern Transvaal, Zimbabwe,

Botswana, Northern Namibia, Southeastern Angola, Southern Zambia, Malawi and

Mozambique (Hogberg, 1986). In Zimbabwe, they are found in the southwest part of

the country (Childes and Walker, 1987).

7

Zambezi teak woodlands occur in sandstone formation in association with P.

Angolensis, Guibourtia coleosperma, Terminalia and Combretum species. Zambezi

teak grows slowly, but is adapted to infertile, deep sands of Kalahari formation

(Bradley and Dewees, 1993). The ‘big three’, in terms of utilisation in Zimbabwe, are

B. plurijuga, P. angolensis and Afzelia quanzensis (Piearce, 1993).

2.3 History of Logging

Exploitation of Baikaea plurijuga started in the indigenous forests of northwestern

Matabeleland in 1907. The officer appointed to oversee logging operations in north

western Matabeleland found that the forests were being overexploited. A forest policy

was then established in the 1920s by the Rhodesia Native Timber Concession, which

harvested teak for railway sleepers (Chenje, Sola and Paleczny, 1998). In subsequent

years, different companies were awarded tenders to extract timber from the forest.

The Forestry Commission granted concessions to companies to perform logging in

Gwayi forest. Between 1996 and 1999, contracts were awarded to the Zimbabwe

Building Services. The contracts were renewable every year. A company called Hyde

Timbers and the local community did logging in 1995. In another area, logging was

last done in 1999 and there has been no logging in another area. (Mashingaidze, pers.

comm)

2.4 Ecological effects of logging

Logging of timber species alters forest ecology by opening up the canopy. Bare

patches created favour the growth of grass and thicket, limiting regeneration through

competition for soil moisture (Bradley and Dewees, 1993). Logging also exposes the

forest floor, thus changing the microclimate of the forest (Calvert, 1986a). For

8

instance, the soil moisture regime is altered by removal of a number of large

dominants, which benefits trees and shrubs that remain. Development of seedlings

and suckers around tree stumps is well pronounced in the opened up areas than in

unlogged forests. This could be due to suppression of undergrowth by forest cover

(Chigwerewe, 1996). Felling of huge trees opens up forest canopy, allowing high

levels of radiation to reach the soil, thus triggering germination of seeds, and

promoting improved growth rate (Chigwerewe, 1996). Logging determines the

pathway of succession. Differences in gap characteristics between logged and

unlogged forests lead to corresponding differences in microclimate, flora, fauna,

frequencies of large herbivore incursions and plant succession (Kasenene, 1987).

There are four major factors responsible for slow post-logging tree regeneration

(Kasanene, 1987, Struhsaker, Lwanga and Kasanene, 1996). These factors are (i)

timber harvesting intensity, (ii) establishment and persistence of an aggressive shrub

or herb layer, (iii) increased elephant use of logged areas compared with lightly

logged and unlogged areas, and (iv) high seed predator rodent densities in the large

gaps of logged areas. Some gaps that are created by logging become infested with

pioneering herbs, vines and shrubs instead of promoting the growth of established

saplings and seedlings (Chapman and Chapman, 1997).

Logging affects species richness. Number of stems of saplings is reduced in the gaps

than in undisturbed sites (Babaasa, et al., 2004). Gaps created by logging do not affect

sapling community composition, species richness or relative species abundance (Uhl,

et al., 1988), and gaps are dominated by shade-tolerant tree species that are present

before gap formation. Gap sizes have an impact on local abundance and species

richness of tree regeneration (Babaasa, et al., 2004). Logging of trees leads to

9

vigorous herbaceous growth (Dye and Spear, 1982; Calvert, 1986a). Successful

regeneration in forests, under both natural and silvicultural variables, depends on

nature of disturbance (Figueroa-Rangel and Olvera-Vargas, 2000).

The vegetation of any given area is dynamic due to intrinsic factors such as

differential longevity of species and environmental factors (Gotosa, 2002).

Disturbance on vegetation can be natural as in the case of drought, floods and fire, or

human induced such as large scale clearing for cultivation or selective harvesting of

plant species like timber logging. Vegetation composition is often influenced by

herbivores through their selective consumption of seedlings and damage to mature

plants (Huntly, 1991). Herbivores also assist with seed dispersal for some plant

species (Wolf and Debussche, 1999).

Lands that have been degraded by previous land use practices such as agriculture and

logging are generally difficult to re-vegetate with desirable species because degraded

soils and competition from undesirable species arrest successional process (Lieth &

Lohman, 1993).

2.5 Logging and fire

Forests in the unlogged areas have the most fire resiliency and present lower fire risk

compared to logged areas (http://www.ems.org/wildfires). This could be due to

limited fuel load since in the logged areas there would be dry stumps that promote fire

establishment. Logging removes the relatively large diameter wood that can be

converted into wood products. This leaves behind small material, which increases the

rate of spread of fires (http://www.ems.org/wildfires).

10

2.6 Canopy gaps

Logging results in the formation of gaps that lead to rapid tree recruitment and

redevelopment of the canopy (Brokaw, 1985a; Denslow, 1987). A forest gap is

created when an open space occurs within the forest and results from the death of a

canopy tree or a broken branch (Timberlake, Muller and Mapaure, unpublished). Gaps

play an important role in plant succession, and hence forest regeneration. It induces

significant changes in the gap microclimate compared to forest understory (Denslow,

1987). Consequently, germination, establishment, growth, and reproduction of many

gap plants are increased (Babaasa, Ellu, Kasangaki, Bifariho and McNellage, 2004).

Canopy gaps associated with selective logging or other major disturbances of tropical

forests increase light reaching the forest floor, and may benefit the growth of the

established saplings and seedlings (Cannon, et al., 1994; Chazdon, 1998). In some

cases, such gaps become infested with aggressive pioneering herbs, vines and shrubs

(Osmaston, 1959; Howlett and Davidson, 1996; Ashton et al., 1997; Chapman and

Chapman, 1997) that appear to repress tree regeneration and halt succession towards a

mature forest community (Sarmiento, 1997; Chapman and Chapman, 1997). At times,

pioneer species are still absent from these sites decades after disturbance (Pinard et

al., 1996; Chapman and Chapman, 1997). Successful regeneration in forests depends

on a disturbance in the canopy that eventually involves gap creation (Figueroa-Rangel

and Olivera-Vargas, 2000).

2.7 Intermediate Disturbance Hypothesis (IDH)

IDH states that diversity will be highest at sites that have had an intermediate

frequency of disturbance, and will be lower at sites that have experienced very high or

11

very low levels of disturbance (Schwilk, Keely and Bond, 1997). This hypothesis is

attributed to Connell (1978) and Huston (1979 as quoted by Schwilk, et al., 1997).

Basically, intermediate levels of disturbance maximise species diversity because

completely dominant species exclude poor competitors at low disturbance but too

much disturbance leads to local extinctions(http://www.findarticles.com/p/articles/mi)

. Essential elements of IDH are that there should repeated local disturbance, creating

bare patches. Furthermore, disturbance must be frequent enough so that competitive

exclusion does not occur over the whole area. Lastly, the frequency of disturbance has

to be seen in the context of generation time (Wilson, 1994).

2.8 Regeneration of B. plurijuga

Natural regeneration in forestry means the renewal of the tree canopy by self-sown

seed or vegetative regrowth (Piearce, 1993). Regeneration is the key process that

influences population dynamics (Begon, Townsend and Harper, 1996). The majority

of Southern Africa’s commercial timber species, their slow growth and the need for

long term management, do not make large scale planting of indigenous trees for wood

products an attractive or economic proposition (Piearce, 1993). In Zimbabwe, the

social forestry fraternity is now learning to build upon local strategies of woodland

management based on natural regeneration (Makuku, 1990).

Teak forests if not grossly abused could be managed by natural regeneration to

replace themselves (DeMeo, 1984). This entails application of the classical selection

system, which entails the creation and maintainace of uneven-aged stands. There are

many factors limiting natural regeneration, and these include erratic seed years, loss

of seed eaten by rodents, and competition for available moisture (Chisumpa, 1984).

12

Regeneration is said to be mainly from seeds, which develops a root that grows up to

0.7-1.0 m during the first rains to about 1.5 m after three seasons (Calvert, 1986a).

This is not consistent with Gambiza’s observations who observed that regeneration of

B. plurijuga was mainly by stem regeneration. An underground wood ‘rhizome’

develops from which new stems arise each season (Calvert, 1986a). Regeneration can

also be by means of coppices, which vigorously grow from stools of any age or size

(Mululuma, 1984). B. plurijuga has a lengthy establishment phase during which the

tape-root develops very extensively, within six months reaching 2 m in length whilst

the stem attains and may remain for several years no taller than 15 cm. The annual

diameter increment for B. plurijuga in Zimbabwe is about 2.5 mm (Calvert, 1986a).

13

CHAPTER THREE

STUDY AREA

3.1 Location

The study was conducted at Gwayi State Forest (Figure 3.1). The forest is located

three kilometres from Lupane Growth Point along the Bulawayo-Victoria Falls road,

northwestern Zimbabwe. It extends from 18˚ 45' S to 19˚ 30' S and from 27˚ 40' E to

28˚ 11' E (Gambiza, 2001).

Figure 3.1: Location of study area in Zimbabwe

14

Figure 3.2: Map showing location of sampling points in the study area

3.2 Geology and soils

The area is underlain by cretaceous sedimentary rocks beneath which are karoo

basalts and sedimentary deposits (The Forest Survey Final Report, 2001). The soils at

the site are deep Kalahari sands that belong to the regosol group in the amorphic soil

order (Nyamapfene, 1991 as cited by Gambiza, 2001). The forest soils are derived

from four main geological formations on parent material (Judge, 1986). The sands are

unconsolidated, red/orange, pink or buff coloured, structureless with a high proportion

of fine dust (Mushove, Gondo and Gumbie, 1993).

15

3.3 Climate

There are three distinct seasons in Gwayi Forest namely: dry winter which stretches

from April to August, hot season which spans October to November and wet season

from mid-November to March (The Forest Survey Final Report, 2001). Rainfall

ranges from 400mm to 800mm per year (The Forest Survey Final Report, 2001) and

is short and erratic with frequent prolonged droughts (Nemarundwe and Mbedzi,

1999). Mean monthly temperature ranges from 15°C (June to September) and 25°C

(October to December) (The Forest Survey Report, 2001).

16

0

50

100

150

200

250

300

350

400

July

Aug

u st

Sep

tembe

r

Oc to b

e r

Nov

embe

r

Dec

embe

r

Janu

a ry

Feb

rua ry

Ma rch

Ap r

il

Ma y

June

Period

Rain

fall (

mm

)

92/9393/9494/9595/9696/9797/9898/99

Figure 3.3: Rainfall pattern in the study area between 1992 and 1999

17

3.4 Vegetation

Zambezi Teak woodlands dominate the site. Vegetation is dominated by Baikiaea

plurijuga commonly associated with Guibourtia coleosperma, Burkea africana,

Erythrophleum africanum, Pterocarpus angolensis, Combretum collinum,Croton

gratissmus, and Schinziophyton rautanenii (The Forest Survey Final Report, 2001).

Other common canopy species are Combretum zeyheri, Commiphora marlothii,

Ochna pulchra, Baphia massaiensis, Terminalia sericea and Bauhinia

petersiana(Gambiza, 2001). Dichapetulum rhodesiana is also common including

Panicum maximum and Heteropogon melanocarpus. Other common grasses are

Pogonarthria fleckii, Aristida stipitata, A. pilgeri, Triraphis schlechteri, Tristachya

rehmanii, Eragrostis species and Digitaria pentzii (Gambiza, 2001).

3.5 Settlements

Settlements were established in Gwayi Forest in the 1930s (Forest Final Survey

Report, 2001). About 882 of the settled families do not have permits and are settled

illegally. The Forestry Commission employs the shared management programme

where the settlers are co-managers of the forest. The families farm, collect forest

products and graze their livestock (The Forestry Survey Final Report, 2001).

3.6 Conservation status

Gwayi Forest was demarcated as state land in 1936 under the Land Apportionment

Act of 1930 (The Forest Survey Final Report, 2001). Before that people lived in the

forests and performed various activities. The Forestry Commission allowed them to

remain in the forest as they assisted with putting out fires (The Forest Survey Final

Report, 2001). The people were issued with permits to remain in the forest. Some

18

settlers moved in from elsewhere. There are four main legal instruments that relate to

wood resources in Zimbabwe. These are the Natural Resources Act of 1941, revised

in 1975 and 1981. This Act created the Natural Resources Board to control use of

natural resources. The second piece of legislation is the Forestry Act of 1949, which

created the Forestry Commission and charged it with control of forest utilisation. The

third piece of legislation is the Communal Lands Acts of 1982, which gives Rural

District Councils the rights to control harvesting of resources in their jurisdiction. The

last piece of legislation the Communal Lands Forest Produce Act of 1987 which

controls the harvesting of listed plant species in communal areas (Gondo, 1993)

19

CHAPTER FOUR

METHODS

4.1 Experimental design, area selection and location of sampling points

A vegetation map of the study area was used to delineate the area into three study

sites based on period when forest was last logged. The three sites were: unlogged area

(control), area logged in 1995 (10 years earlier prior to the study) and area logged in

1999 (four years earlier prior to the study). This delineation was based on Forestry

Commission’s logging records. The study was based on a completely randomised

design, with three treatment levels, namely unlogged, logged four years earlier

(1999), and logged 10 years earlier prior to the study (1995). Sampling points were

selected in each sampling site by super-imposing a grid system on an aerial

photograph. The sampling points were located on the grid intercepts. Fifteen sampling

points were randomly selected from each of the three sampling sites. A

reconnaissance, followed by ground truthing was conducted at the beginning of

March 2005. This had the purpose of ascertaining area boundaries, and confirming

presence or absence of the tree species of interest.

Sampling commenced immediately after ground truthing in late March 2005.

Selection criteria of sampling points were that the sampling point should have been

dominated by B. plurijuga, with at least 15 to 20 plants along a 50 m long stretch,

25m on either side of the point. The sampling point should not have been burned since

logging, hence charred and burnt logs were searched for as a means of verification.

Sampling points, which did not meet these criteria, were discarded. The distances of

sampling points from the Victoria Falls-Bulawayo road were estimated on the aerial

photograph using strings to trace distances and map scales to estimate distances. They

20

were then located in the field by estimating distance from the main road to the

sampling point with the assistance of a Global Positioning System instrument.

4.2 Demarcation of plots

At each sampling point, a 50-m string was laid so that half of the string was on either

side of the sampling point. Fifteen to twenty individuals of B. plurijuga were then

counted on either side of the stretched string (Walker, 1976). On getting this number

of trees, counting stopped as such a point would qualify to be assessed. Assessment

was carried out within the plot. Plots were, therefore, parallel to the 50m stretches,

and formed the main plot. This produced plots of fixed length but variable width.

Simultaneously, three sub-plots (measuring 5m x 5m) were randomly placed within

the main sampling plot for the assessments of saplings. Three smaller subplots

(measuring 1m x 1m) were randomly placed in each sapling subplot for the

assessment of seedlings.

4.3 Measurement and assessment of woodland structure

4.3.1 Plot counts

For each plot, B. plurijuga individuals were assessed for basal diameter and height,

whilst other tree species were merely identified and counted. In this study, trees were

defined as plants with a basal diameter of 0.060 m or more, and a height of 3 m or

more. Saplings were defined as woody plants with a basal diameter of between 0.01 m

and 0.059 m, but less than 3 m in height. Individuals less than 1 m in height were

regarded as seedlings. In order to be included in the assessment, at least half the

rooted stem of the plant had to lie within the plot in the case of multiple-stemmed

shrubs, and live foliage lying within the plot was measured (Guy, 1981). Unless

21

multi-stemmed trees were distinctly joined at the base, the stems were measured

separately. Multi-stemmed individuals with an average stem diameter of 0.06 m or

more were regarded as trees, and each stem was recorded separately (Guy, 1981).

Seedling heights were measured using a ruler, whilst those for other plant species

were only identified and counted. Species that could not be identified were given

identity letters, enumerated, their specimens collected and pressed, and taken to the

National Herbarium for later identification.

4.3.2 Measurement of basal diameter and height

A tape measure was used to measure stem circumference of B. plurijuga 0.3 m above

the base of the stem. The data was used to calculate the basal area using, ba = (C)2/4π

where C stands for basal circumference and ba stands for basal area. For multi-

stemmed trees, the basal circumference of each stem was measured separately and

total basal area calculated. The heights of B. plurijuga were estimated by use of

ranging rods. The data was entered in a field data sheet, which recorded logging state,

B. plurijuga tree height, circumference, basal area and other tree species

4.3.3 Assessment of saplings and seedlings

The number of saplings was counted in three sub-plots each measuring 5m x 5m

located at random within each tree-sampling plot. The height of each sapling was

estimated using a ranging rod. Basal circumference was measured with a tape

measure. The assessment of seedlings was done on three sub-plots measuring 1m x

1m located at random within each sapling-sampling plot. The height of each seedling

was measured with a ruler from the base of the stem to the apical end. Data were

recorded on field data sheets.

22

4.3.4 Assessment of stumps

Stumps of B. plurijuga were identified and counted in each plot. Each stump’s status

(with or without coppices) was recorded. Basal circumference was measured using a

tape measure. Stumps represented clear evidence of logging.

4.4 Measurement of explanatory variables

4.4.1 Soil sampling

Soil samples were collected from three random points in the main plot. At each point,

plant litter was removed prior to collection of the soil sub-sample. Soil samples were

collected using a soil augur. Sub-samples were collected from a depth of 0.60 m. Soils

from each plot were bulked to form a composite sample, and bagged for later

determination of pH, nitrogen, phosphorus and potassium at the Soil Science and

Engineering laboratory, University of Zimbabwe.

4.5 Data analyses

Hierarchical Cluster Analysis (HCA) (MINITAB, 2000) using average linkage

method was performed on a matrix of 45 plots by 48 species to generate clusters of

sample plots based on species presence/absence data. Cluster analysis develops

meaningful aggregations, or groups, of entities based on a large number of

independent variables (McGarigal, Cushman and Stafford, 2000). Its objective is to

classify a sample of entities into a smaller number of usually mutually exclusive

groups based on the multivariate similarities among entities (McGarigal, et al., 2000).

Average linkage is at times called unweighted pair-group average and it designates

distance values between groups to be the average dissimilarity between clusters.

When two clusters agglomerate, their dissimilarity is equal to the mean of the

23

distances between each entity in one cluster with each entity in the other cluster. The

entities or clusters that have the lowest mean distance value detect fusion. This

algorithm maximises the correlation between input dissimilarities and output

dissimilarities implied by the resulting dendrogram (McGarigal, et al., 2000).

Detrended Correspondence Analysis (DCA) (ter Braak, 1986, 1995, Gauch, 1982)

was performed on species presence/absence data to elucidate relationships amongst

the various plant associations and underlying environmental gradients (Mapaure,

2001). DCA is a reciprocal averaging technique, whose objective is to remove the

arch effect and to compress the axis (McGarigal, et al., 2000). ‘Detrending’ to

eliminate the arch effect is accomplished by dividing the first axis into a number of

equal segments, and within each segment, adjusting the ordination scores to a mean of

zero. DCA has the advantage of being able to handle large, complex data sets, and

uncovering extremely long ecological gradients, and has no arch effect. This

technique performs well when data have non-linear and unimodal distribution

(McGarigal, et al., 2000). Its most important limitation is its sensitivity to outliers and

discontinuities in the data, and its poor performance with skewed variable

distributions (Palmer, 1993).

Canonical Correspondence Analysis (CCA) (ter Braak, 1987, 1988) contained in the

programme package CANOCO version 4.0 (ter Braak, 1998) was used to explore

species-environment relationships on the same species data set subjected to DCA. In

this analysis, four explanatory variables were used. CCA is a hybrid of ordination and

multiple regression (McGarigal, et al., 2000). It arranges species along environmental

variables and constructs those linear combinations of environmental variables, along

24

which the distributions of species are maximally separated (ter Braak, 1987). In this

technique, ordination axes are constrained to linear combinations of environmental

variables. The technique is used for detecting species–environment relationship, and

for investigating specific questions about the response of species to environmental

variables (ter Braak, 1987). In the CCA ‘triplot’, the distribution of species and

sample points jointly represent the dominant ecological relationships as they are

explained by explanatory variables (McGarigal, et al., 20000. The explanatory

variables from the second matrix are plotted in the CCA triplots as arrows emanating

from the grand mean of all explanatory variables (ter Braak, 1986). The arrows in the

ordination space indicate the direction of maximum change in each structuring

variable with its length being equal to the rate of change of the weighted averages.

CCA provides a measure of species distributions change along that explanatory

variable, and the length indicates the importance of an environmental variable

(McGarigal, et al., 2000). The location of species points relative to the arrows

indicates the characteristics of the ecological optima of each species (variable

(McGarigal, et al., 2000). The significance of the relationship between floristic and

environmental variables was tested with the Monte Carlo permutation test, which

contained within programme package CANOCO. This is a test of significance

obtained by repeatedly shuffling (permuting) the sample (ter Braak and Smilauer,

1988).

Statistical significance of differences in the size class distribution patterns of the

species among three areas was carried out using Chi-squared test. Size class structure

of B. plurijuga was explored through the analysis of size class data. This gives an

indication of recruitment at any one particular stage in the population’s history.

25

Data for species counts was used to calculate the Shannon-Weiner’s Diversity index

(H΄) (Ludwig and Reynolds, 1988) for each plot. These indices were then tested for

normality using Kolmogorov-Smirnov test and were found to be normal. The indices

were subjected to One-way ANOVA found in programme package MINITAB

(MINITAB, 2000) to elucidate differences in diversity for each plot with logging

status as a factor. There were areas that had never been logged, areas that were logged

four years earlier, and areas that were logged ten years earlier. Treatments were: never

logged (1), logged in 1995 (2) logged in 1999 (3). The Shannon-Weiner’s Diversity

index was used because it is simple to use. It is calculated by: H' = -Σ (pi) (ln pi),

where H' is the index of species diversity, pi = proportional abundance of species ln,

and I = natural logarithm (Krebs, 1972). This index incorporates both species richness

and evenness into a single value (Ludwig and Reynolds, 1988). It is a measure of

‘uncertainty’ in predicting to what species one individual chosen at random from a

collection will belong (Ludwig and Reynolds, 1988). This ‘uncertainty’ increases as

the number of species increases, and as the distribution of individuals among species

becomes even (Ludwig and Reynolds, 1988). Diversity indices can be used to

evaluate the intensity of resources used by human populations, and can also help to

determine a minimum area necessary for a native population based on data on

resources used (Begossi, 1996). The Shannon-Weiner’s Diversity index is regarded as

having a moderate sensitivity to sample size (Maguran, 1988).

Data for explanatory variables, namely pH, N, P and K were tested for normality

using the Kolmogorov-Smirnov test, and were found to be normally distributed. There

26

was no need for transformation of the data. Data were subjected to One-way ANOVA

with each environmental variable as factors and logging period as treatments.

Data for tree density and stump density were tested for normality using the

Kolmogorov- Smirnov test and was found to be not normal. Log-transformation using

Log10 (x + 1) was tried prior to testing the data for normality again using the above

test. The data still could not pass the normality test. Due to failure of transformation, a

non-parametric procedure was then employed. The significance of the two variables

was then tested using Kruskal-Wallis test.

27

CHAPTER FIVE

RESULTS

5.1 Species composition and population structure of B. plurijuga in the sampling areas A total of 824 woody species in the unlogged area, 549 in the area logged in 1995 and

603 in the area logged in 1999 were recorded (Appendix 23). Two hundred and forty

three individuals of B. plurijuga were encountered in the unlogged area, 239 in the

area logged in 1999 and 244 in the area logged in 1995 (Appendix 6). Size class

distribution followed almost the same overall pattern in the unlogged area, area

logged in 1995 and area logged in 1999 (Figure 5.1, Figure 5.2, Figure 5.3 and Figure

5.4). The graphs show roughly an inverse ‘J size class distribution’ with the greatest

proportion of individuals in the lower diameter size class. The only difference was the

greater proportion (and absolute densities) of individuals (size class <0.100 m) in the

area logged in 1995 than in the unlogged area and in the area logged in 1999 (Figure

5.4). No B. plurijuga trees were recorded in the 0.501-0.600 m and 0.601-0.700 m

classes in the unlogged area, area logged in 1995 and area logged in 1999 (Figure 5.1,

Figure 5.2 and Figure 5.3). In the unlogged area there was a small proportion of trees

in the 0.701-0.800m diameter size class but they was none in the area logged in 1999

and in the area logged in 1995 (Figure 5.1, Figure 5.2 and Figure 5.3). Results

indicate that the highest proportion of trees in the 0.301-0.400 m diameter class was in

the unlogged area (Figure 5.4). Basal diameter size classes were significantly different

in the unlogged area, area logged in 1999 and area logged in 1995 (X2 = 20.721, df =

10, p< 0.05) (Appendix 16).

28

0

10

20

30

40

50

60

70

< 0.100 0.101-0.200 0.201-0.300 0.301-0.400 0.401-0.500 0.501-0.600 0.601-0.700 0.701-0.800

Basal diameter class (m)

Per

cent

age

of to

tal p

opul

atio

n

Figure 5.1: Basal diameter classes of B. plurijuga in an area logged in 1995 in Gwayi Forest

29

0

10

20

30

40

50

60

< 0.100 0.101-0.200 0.201-0.300 0.301-0.400 0.401-0.500 0.501-0.600 0.601-0.700 0.701-0.800

Basal diameter class (m)

Perc

enta

ge o

f tot

al p

opul

atio

n

Figure 5.2: Basal diameter classes of B. plurijuga in an unlogged area in Gwayi Forest

30

0

10

20

30

40

50

60

< 0.100 0.101-0.200 0.201-0.300 0.301-0.400 0.401-0.500 0.501-0.600 0.601-0.700 0.701-0.800

Basal diameter class (m)

Perc

enta

ge o

f tot

al p

opul

atio

n Figure 5.3: Basal diameter size classes in an area logged in 1999 in Gwayi Forest

31

< 0.1000.101-0.200

0.201-0.3000.301-0.400

0.401-0.5000.501-0.600

0.601-0.7000.701-0.800

Area 1

Area 2

Area 30

10

20

30

40

50

60

Perc

ent o

f tot

al p

opul

atio

n

Basal diameter class (m)

Area 1Area 2Area 3

(Area 1 represents unlogged area, Area 2 represents area logged in 1995 and Area 3 represents area logged in 1999) Figure 5.4: Basal diameter classes in the unlogged area, area logged in 1995 and area logged in 1999 in Gwayi Forest

32

5.2 Species richness and species diversity

Data for calculating richness and diversity is in Appendix 23. Highest richness of 11

was found in Plots 4 and 5 found in an unlogged area and Plot 30 found in an area

logged in 1995 (Appendix 13). The least richness was found in Plot 29 found in an

area logged in 1995 and Plot 45 found in area logged in 1999. Highest mean richness

was in the unlogged area and in the area logged in 1999 (Figure 5.6). Highest

diversity was found in Plot 6 found in an unlogged area whilst lowest diversity was

found in Plot 29 located in an area logged in 1995 and plot 45 located in area logged

in 1999 (Appendix 13). Mean diversity was highest in the unlogged area and lowest in

the area logged in 1995 (Figure 5.5). Results show that richness was highest in the

area logged in 1995 and lowest in the area logged in 1999 (Appendix 23). There was

no significant difference in species diversity, species richness, basal area, seedling

regeneration and sapling regeneration (F=1.37, P>0.05; F=0.023, p>0.05; F=2.38,

p>0.05; F = 0.47, p > 0.05; F=0.84, p>0.05) (Appendices 5, 9, 10, 21 and 22)

respectively among the three sampling areas.

Table 5.1: Species richness and diversity in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest

Area Richness Diversity (H’) Unlogged 29 2.39

Logged in 1999 28 2.26

Logged in 1995 33 1.90

33

Area

0 1 2 3 4

Mea

n dive

rsity

inde

x (H

')

0.0

0.5

1.0

1.5

2.0

2.5

Figure 5.5: Differences in mean diversity in an unlogged area, area logged in 1999 and an area logged in 1999

Area

0 1 2 3

Mea

n ric

hnes

s

0

2

4

6

8

10

12

4

Figure 5.6: Differences in mean species richness in an unlogged area, areas logged in 1995 and 1999 (Area 1 represents the unlogged area, Area 2 logged in 1995 and Area 3 logged in 1999)

34

5.3 B. plurijuga densities, basal area, extent of logging and regeneration

Tree density was highest in the unlogged area. Density was similar in areas logged in

1999 and 1995 (Appendix 13). Results indicate that the tree and stump densities of B.

plurijuga trees between the unlogged area, area logged in 1995 and area logged in

1999 were not significantly different (H = 43.27, p>0.05; H = 0.19, p> 0.05) Table

5.4, Appendix 7 and 8) respectively. Total basal area was highest in the unlogged area

and lowest in the area logged in 1999 (Table 5.3). Mean basal area was highest in the

unlogged area and lowest in the area logged in 1999 (Figure 5.7). Stump density was

higher in the area logged in 1999 than in the area logged in 1995 (Table 5.2). Mean

stump density was higher in the area logged in 1999 than in the area logged in 1995.

Results indicate that more stumps were coppicing in the area logged in 1999 than in

the area logged in 1995 (Table 5.5, Appendix 12). The density of seedlings was

highest in the area logged in 1999 and lowest in the area logged in 1995 (Table 5.5).

Mean number of seedlings was highest in the area logged in 1995 and lowest in the

area logged in 1999 (Figure 5.9). The density of saplings was highest in the area

logged in 1999 and lowest in the unlogged area (Table 5.5). The mean number of

saplings was highest in the area logged in 1995 than in the area logged in 1999

(Figure 5.10). The highest number of seedlings, saplings and trees were recorded in

the area logged in 1995 and the lowest was recorded in the area logged in 1999

(Figure 5.11). Approximately the same percentage of seedlings (<0.01m) that were

recruited into the saplings basal diameter class (0.01-0.599 m) in the unlogged area

was the same as the percentage number of saplings that were recruited into the tree

basal diameter class (> 0.06m). In the area logged in 1995, fewer seedlings recruited

into the saplings diameter class. More seedlings recruited into the saplings diameter

35

class in the area logged in 1999. In the unlogged area, more saplings were recruited

into the tree basal diameter class. In the area logged in 1995, fewer saplings were

recruited into the tree basal diameter class. In the area logged in 1999, more

individuals were recruited into the tree basal diameter class.

Table 5.2: Density of B. plurijuga and extent of logging in an unlogged area, area logged in 1999 and area logged in 1995 Area Tree density/ha Stump density/ha

Unlogged 0.033 0

Logged in 1999 0.004 0.021

Logged in 1995 0.004 0.020

Table 5.3: Total basal areas in square metres for B. plurijuga in unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest

Area Total basal area in square metres Unlogged area 6.729

Area logged in 1999 4.5144

Area logged in 1995 5.4631

36

Table 5.4: Summary of the Kruskal-Wallis test results of tree and stem densities in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest

Areas Variable H-value p-value df

1, 2 and 3 Tree density 43.27 0.460 43 (adj)

Tree density 38.48 0.668 43

2 and 3 Stump density 0.19 0.663 1

(Area 1 represents unlogged, Area 2 represents logged in 1995 and Area 3 represents

logged in 1999)

Area

0 1 2 3

Mea

n ba

sal a

rea

0.0

0.2

0.4

0.6

0.8

1.0

4

Figure 5.7: Differences in mean basal area in an unlogged area, Area logged in 1999 and Area logged in 1995 in Gwayi Forest Area 1 represents an unlogged area, Area 2 represents the area logged in 1995 and Area 3 represents area logged in 1999)

37

Area

0 1 2

Mea

n st

ump

basa

l are

a (s

qu. m

etre

s)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

3

Figure 5.8: Differences in mean stump densities in an area logged in 1999 and an area logged in 1995 (Area 1 represents area logged in 1995 and Area 2 logged in 1999) Table 5.5: Regeneration of B. plurijuga in an unlogged area, area logged in 1999 and area logged in 1995

Area % Stumps coppicing Seedlings/ha Saplings/ha Unlogged 0 0.042 0.045

Logged in 1999 57.4 0.0610 0.061

Logged in 1995 22 0.0414 0.047

38

Area

0 1 2 3

Mea

n nu

mbe

r of s

eedl

ings

0

1

2

3

4

5

4

Figure 5.9: Mean number of seedlings in an unlogged area, area logged in 1999 and an area logged in 1995 (Area 1 represents the unlogged area, Area 2 logged in 1995 and Area 3 logged in 1999)

39

Area

0 1 2 3 4

Mea

n nu

mbe

r of s

aplin

gs

0

1

2

3

4

5

6

Figure 5.10: Differences in mean number of saplings in the logged area, area logged in 1999 and area logged in 1995 in Gwayi Forest (Area 1 represents an unlogged area, Area 2 logged in 1995 and Area 3 logged in 1999)

5.4 Vegetation classification

Presence/absence data used in this analysis is in Appendix 24. Based on species

presence/absence data, seven floristic associations can be identified on the

dendrogram (Figure 5.11). The species presence/absence matrix used in the

multivariate analysis is presented in Appendix 24. Three plots from the unlogged area

were grouped into one cluster with more than 65% similarity. These plots were

dominated by B. plurijuga. The other common tree species were: Terminalia sericea

and Pterocarpus angolensis. The plots were found in an area that is characterised by

deep, brown well-drained sandy soils. Cluster 2 had seven plots, all from the unlogged

area. The dominant tree in these plots was B. plurijuga. Other common trees species

were Croton gratissimus and Commiphora mollis. The plots were found in an area

with sandy, deep, brown and well- drained soils. The third cluster comprised five

plots, two coming from an area logged in 1995 and three from an unlogged area. The

40

common tree species was Croton gratissimus. Cluster 4 had two plots, both from an

area logged in 1995. Dominant tree species is B. plurijuga. Other common tree

species are Commiphora mollis, Acacia species, Combretum apiculatum and

Combretum collinum. The soils were deep, brown and sandy. Cluster 5 had eight

plots, two from an area logged in 1999 and six from an area logged in 1995. B.

plurijuga dominates these plots. Commiphora mollis was the common tree species in

these plots. Cluster 6 had fifteen plots, two from the unlogged area, and thirteen from

an area logged in 1999. The plots had brown, deep and well-drained soils. The

dominant tree was B. plurijuga. Other common trees were: Strychnos pungens,

Diplorynchus condylocarpon, Combretum collinum, Terminalia sericea and

Combretum molle. Cluster 7 comprised five plots, all from the area logged in 1995.

These plots were found in a gentle sloping area, with deep and well-drained soils. B.

plurijuga was the dominant tree. Other common trees were Grewia rertinevis,

Commiphora mollis and Combretum molle.

41

C 1 C 2 C3 C 4 C5 C6

C7

C 1 C 2 C 3 C 4 C 5 C 6 C 7

3 4 5 6 8 9 7 11

12

10

13

15

22

14

21

26

29

23

43

24

30

27

45

25

28 2 31

32

34

35

36

38 1 41

33

37

39

44

40

42

16

17

18

20

19

100.00

87.32

74.64

61.97

Plot number

Sim

ilarit

y (%

)

Figure 5.11: Hierarchical Cluster Analysis dendrogram showing classification of vegetation plots based on species presence/ absence data (C represents cluster, for example, C 3 represents cluster 3)

5.5 Species–environmental relationships

Presence/absence data for this analysis is in Appendix 24. Results of Canonical

Correspondence Analysis (CCA) applied to four environmental variables and forty

five plots indicates that that axis 1 accounted for 23.2% of observed variation,

whereas Axes 2, 3 and 4 account for 12.8%, 10% and 6.7%, respectively (Table 5.6

and Appendix 18). The forty-five plots were separated by visual inspection into three

groups. Group 1 (18, 21 and 23) are found in an area logged in 1995 and common

42

species are Combretum molle and Combretum zeyheri. The dominant species in all the

forty-five plots was B. plurijuga. Group 2 (44, 43, 40, 35, 34, 36, 32, 24, 41, 28, 45,

39, 37, 38, 42 and 31) were found in an area logged in 1999. Common species were:

Terminalia sericea, Strychnos pungens, Diplorhynchus condylocarpon, Combretum

collinum and Combretum molle (Figure 5.13). Group 3 (26, 33, 27, 19, 16, 20, 30, 25,

22, 29 and 17) were found in an area logged in 1995 and (11, 3, 7, 5, 4, 6, 10, 1, 2, 15,

13, 8, 12, 14 and 9) were found in an unlogged area. Common species were:

Commiphora mollis, Acacia species, Terminalia sericea and Combretum collinum.

The analysis showed that nitrogen was positively correlated with axis 1 (r = 0.54) and

phosphorus positively correlated with axis 2 (r = 0.16). pH was negatively correlated

with axis 1(r = -0.80) and potassium was positively correlated with axis 2 (0.02)

(Appendix 18). pH was negatively correlated with phosphorus (r = -0.10) and nitrogen

(r = -0.49) (Appendix 18). Phosphorus was positively correlated with Potassium (r = -

0.09) and negatively correlated nitrogen (r = -0.01) (Appendix 18). Nitrogen was

negatively correlated with Potassium (r = -0.24). Out of the 4 environmental variables

involved in the analysis, Phosphorus (F = 1.72, p <0.05) and pH (F = 2.50, p < 0.05)

contributed significantly to the total variance of floristic data (Appendix 18). Arrows

in Figure 5.13 show the environmental gradients, their relative importance and

intercorrelation of environmental variables.

5.6 Indirect gradient analysis

In the DCA ordination including all forty-five plots, species composition

corresponded best to the first axis than the second DCA axis. The total variation in

species data explained along the first DCA ordination was 34% while 24.8 % was

explained along the second (Appendix 19 and Table 5.6). Plots were separated into

43

three main recognisable groupings according to species composition. Detrended

Correspondence Analysis showed clustering of plots that have never been logged,

those that were logged in 1995 and those that had been logged in 1999. The first axis

seems to be associated with phosphorus variation. The eigen values for the first two

axes of the Detrended Correspondence Analysis were 0.34 and 0.248, respectively,

and explained 15.5 % of species variance. Group 1 represents B. plurijuga,

Combretum collinum and Terminalia sericea. Group 2 represents B. plurijuga,

Commiphora mollis, Terminalia sericea, Strychnos spinosa and Diplorrhynchus

condylocarpon. Group 3 represents B. plurijuga, Commiphora mollis, Terminalia

sericea, Combretum collinum, Grewia monticola and Grewia retinervis.

44

Axis 1

Axi 2

-1.5 +2.0

-1.5

+2

pH

N

P

K

p21

p38

p23

p37 p41

p40 p34

p35

p39

p4

p43

p36

p9

p33

p18

p 17

p5

p45

p44

p2 p3

p6

p14

p11

p10

p13

p8p7

p12

p1

p31 p32

p42

p29

p19

p15p22

p27

p16

p30p24

p28 p26

p20

p25

(Axi 2 represents Axis 2) Figure 5.12: CCA ordination plot showing influence of edaphic factors on the distribution of sample plots in Gwayi Forest

45

+5.5 36 34 17 16 18 30 19 31 1 3 37

Axis 2 42 29 20 26 27 5 11 14 33 38 44 7 2 8 9 25 24 12 22 23 21 15 10 32 28 41 40 4 39 13

35 Axis 1 +8

43

--1

--1

9

30

8 38

3

2

1

Figure 5.13: Position of forty five plots on the first two axes of DCA analysis

46

Table 5.6: Eigenvalues and percentage contribution of each component in the DCA and CCA analysis ordination of plots in Gwayi Forest

Dimension Eigenvalue Percentage contribution (%) Cumulative percentage

(%) CCA

1 0.232 23.2 44

2 0.128 12.8 68.4

3 0.100 10 87.3

4 0.067 6.7 100

DCA

1 0.34 34

2 0.248 24.8

3 0.174 17.4

4 0.132 13.2

5.8 Influence of edaphic factors on vegetation distribution

Results show that mean Nitrogen (ppm) was highest in the unlogged area, and lowest

in the area logged in 1999 (Figure 5.15). Differences in Nitrogen (ppm) were

significant between the unlogged area, area logged in 1999 and area logged in 1995 (F

= 3.07, p< 0.05, Appendix 4). Mean pH was highest in the area logged in 1999 and

lowest in the area unlogged area (Figure 5.18). Differences in pH were significantly

different in the unlogged area, area logged in 1999 and area logged in 1995 (F =9.20,

p < 0.05) (Appendix 1). Mean Phosphorus (ppm) was highest in the area logged in

1995 and lowest in the area logged in 1999 (Figure 5.17). Differences in soil

Phosphorus (ppm) were not significant in the unlogged area, area logged in 1999 and

area logged in 1995 (F =0.47, p > 0.05, Appendix 2). Mean Potassium (me %) was

highest in the area logged in 1995 and lowest in the unlogged area (Figure 5.16). Soil

47

Potassium (me%) of the unlogged area, area logged in 1999 and area logged in 1995

was not significantly different (F = 0.85, p > 0.05, Appendix 3).

Table 5.7 Summary of the significance of the effects of area variables on species composition in Gwayi Forest

Variable F-ratio p-value Nitrogen (N) 3.07 0.006 Potassium (K) 0.85 0.556 Phosphorus (P) 0.47 0.850 pH 9.20 0.000

Area

0 1 2 3

Mea

n N

itrog

en (p

pm)

0

2

4

6

8

10

12

14

16

4

Figure 5.14: Differences in mean Nitrogen (ppm) in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest

48

Area

0 1 2 3

Pot

assi

um (m

e %

)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

4

Figure 5.15: Differences in mean potassium (me%) in an unlogged area, area logged in 1999 and in an area logged in 1995 in Gwayi Forest

Area

0 1 2 3

Mea

n Ph

osph

orus

(ppm

)

0

5

10

15

20

25

4

Figure 5.16: Differences in mean Phosphorus (ppm) content in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest

49

Area

0 1 2 3 4

pH

0

2

4

6

8

Figure 5.17: Differences in mean pH in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest

50

CHAPTER SIX

DISCUSSION

6.1 Population structure

An ‘inverse J’ size class distributions in the unlogged area, area logged in 1995 and

area logged in 1999 suggests that logging did not affect size class. This represents a

stable distribution. This could indicate the existence of a cohort of immature trees that

may have experienced a disturbance (Campbell, 1988). An irregular size class

distribution on the other hand, would suggest the presence of episodic recruitment and

irregular growth patterns (Botha, Witkowski and Shackleton, 2000). The ‘inverse J’

shapes also gives an idea of the recovery potential of each area. It is possible that the

areas logged in 1999 and 1995 were recovering from logging as indicated by the

prevalence of B. plurijuga in lower size classes. This type of distribution suggests

continuous regeneration (Judge, 1986) and possible recruitment of seedlings in the

study areas.

The absence of trees in the 0.501- 0.600 m and 0.601-0.700 m diameter size class in

the unlogged area cannot be explained by these results since one would expect to

record trees in all diameter classes. The higher prevalence of individuals in the 0.301-

0.400 m basal diameter class in the unlogged area when compared to the logged areas

suggest that this could be the basal diameter size class that is targeted for during

logging, hence its prevalence could be attributed to absence of logging. The presence

of trees in the 0.701 m -0.800 m diameter class further confirms the absence of

logging or significant disturbance in the unlogged area.

51

Significant differences in basal diameter size classes suggests that recruitment of B.

plurijuga individuals from a lower size class to another was taking place at different

rates in the unlogged area, area logged in 1999 and area logged in 1995. Bormann

and Likens (1979) concluded that clear felling of a forest is a drastic disturbance that

leads to rapid successional changes. The high prevalence of B. plurijuga in the

seedling and sapling basal diameter size classes could suggest adequate regeneration

through propagules in all the three sites. Higher densities of seedlings and saplings

compared to that of larger trees could have been due to higher initial seedling and

sapling recruitment and frequent deaths, which decline with time (Swaine and

Liberman, 1987).

6.2 Diversity and richness

Diversity was not significantly different among the unlogged area, area logged in

1999 and area logged in 1995. This falsifies the alternative hypothesis that diversity

was significantly different in the study areas. Some of the plots showing high

diversity values were found in the unlogged area. The unlogged area could have had

highest diversity since no trees were removed through logging. Herbivory was also

not intense. Probably the unlogged area had species that were shade–tolerant and were

not suppressed by shade. The findings are not consistent with the conclusion that

vegetation removal leads to rapid growth which leads to increased species diversity

(Kobayashi, Hori and Nomoto, 1977). Were this conclusion true, then highest mean

diversity could have been in the logged areas and not in the unlogged area.

Lowest species richness in the area logged in 1999 could mean that the vegetation

change was not yet complete or there was a permanent loss of some species caused by

52

the disturbance (Gotosa, 2002). Highest mean richness in the unlogged area could

suggest that the canopy tree species did not suppress other tree species or the woody

species growing there were shade-tolerant. Species richness in forests was also

observed to be low. Species richness is reported to decline with increasing stand

disturbance (Sabogal, 1992). This observation is consistent with the results obtained

in this research.

6.3 Density and basal area

B. plurijuga in the unlogged area could have inhibited the establishment of other

species (Connell and Slatyer, 1977). Higher stump density in the area logged in 1999

could suggest slightly higher logging intensity compared to the area logged in 1995.

Alternatively, it could have been due to logging by timber poachers since the area

logged in 1999 is closer to Lupane Growth Point where demand for timber and timber

products is high. Highest basal area recorded in the unlogged area further confirms

that logging was a major disturbance, which reduced the basal area. Lowest basal area

in the area logged in 1999 confirms that logging could have been more intense than in

1995.

6.4 Regeneration of B. plurijuga

Differences in stump regeneration with higher numbers of coppicing stumps in the

area logged in 1999 than in the area logged in 1995 could suggest that some of the

stumps that were logged in 1995 could have died as a result of cutting and inadequate

moisture in the 1995/1996 rain season (Gambiza, 2001). Moisture is critical for

regeneration. Any damage to a mature stem has been observed to inhibit future

growth due to physiological stress (Tschaplinski and Blake, 1989; Baillie and Ashton,

53

1983). Furthermore, resprouting ability of trees is constrained by anatomy and

physiology of the buds in stumps that are poorly cut (Paciorek, Condit, Hubbell and

Foster, 2000). In 1995/1996 season, there was low rainfall. This could have resulted

in the death of some stumps since regeneration is negatively correlated with shortage

of moisture.

Differences in seedling and sapling abundance, with highest saplings and seedlings in

the area logged in 1999 compared to other sites further confirm that adequate

moisture is critical for regeneration. The results are consistent with findings that when

a gap is created, suppressed seedlings are released and higher light levels are

experienced in the gap (Ng, 1978; Oberbaur and Strain, 1985). Saplings and seedlings

present on the site before gap creation may take advantage of the gap to grow towards

the upper canopy of the forest (Ng, 1978; Oberbaur and Strain, 1985). De Steven

(1991) showed that seedling survival and growth in early succession is selective due

to influence of environmental factors and herbivore selection. Locally, the density of

seedlings and saplings of B. plurijuga were observed to be high in the opened up areas

than in unlogged areas (Chigwerewe, 1996). In the unlogged area, the establishment

of seedlings could have been inhibited by the presence of large canopy trees, which

limited the amount of light and nutrients reaching the soil. Furthermore, below

average rainfall in 1994/1995 season could have resulted in the death of seedlings

(Gambiza, 2001). Availability of moisture has been shown to affect seedling

establishment (Walker, 1987).

54

6.5 Species associations

Plots separated according to the period of logging suggesting the importance of

logging in forest structure. Cutting of wood has been observed to quickly bring

irreversible changes in species composition, physiognomy and structure of vegetation

leading to replacement of vegetation by dense, low thicket (Liberman and

Mingguang, 1992). The patterns observed could have been due to local variations in

soil mineral content. Local variations in edaphic conditions involving features such as

soil depth and soil reaction have been observed to strongly influence the distributions

of many woodland species and result in the development of mosaics (Packham and

Harding, 1982). The association between plots from the unlogged area, plots from an

area logged in 1999 and in 1995 from hierarchical cluster analysis, DCA and CCA

could suggest that there was recovery of vegetation from logging.

6.6 Effect of pH, Nitrogen, Phosphorus and Potassium on Species distribution

Significant differences in pH with the highest values in the area logged in 1999 cannot

be explained by these results. With the removal of trees resulting in an increase in

leaching, the soils could have become acidic. Low pH in the unlogged area could have

been due to the abundance of organic matter leading to production of organic acids by

microorganisms. The results are consistent with the conclusions by (Escudero,

Garrido, Matias and Del Arco, 1988) that open woodlands cause strong spatial

heterogeneity in environmental conditions due to shading and soil enrichment by tree

crowns.

Significant differences in soil Nitrogen among the sampling areas with high values in

the unlogged area could suggest that debris left after logging increased the amount of

55

organic matter which was later converted to ammonium compounds by micro

organisms. Low nitrogen values in the area logged in 1999 suggest that logging

reduced the amount of plant litter that fell into the ground. Consequently, the amount

of organic material converted to ammonium compounds, and ultimately nitrogen,

could have been reduced.

Highest mean potassium in the area logged in 1995 could be due to less leaching since

vegetation recovery could have taken place. High potassium values were expected to

be low in logged areas since leaching of highly soluble potassium could have been

much more pronounced in the area logged in 1995 due to absence of canopy cover.

Generally, the amount of soil potassium was significantly low since fertility is low in

the Kalahari sands (Gambiza, 2001).

Phosphorus values were low according to the Bray method (Moukam and Nyakanou,

1997). This could have been due to intensity of land use in the area. In the unlogged

area, grazing is a major form of land use, and this could affect phosphorus content.

The highest phosphorus values in the area logged in 1995 could be due to the

recovery of vegetation after logging. The lowest phosphorus values recorded in the

area logged in 1999 could be a result of more intense logging. Low mean values of

phosphorus could reflect high intensity of past land use (Moukam and Nyakanou,

1997). Zimbabwean soils are generally deficient in P, which explains why farmers are

encouraged to supplement P to farm animals during the rainy season when much of it

is leached away.

56

CONCLUSIONS

B plurijuga is exploited at basal diameter of between 0.3m and 0.4m for commercial

use. At current rates of exploitation of B plurijuga in Gwayi forest, logging has no

effect on size class distribution, species diversity, species richness, regeneration

potential (through coppicing or seedlings) and stump density. This is explained by

intensity of logging that is light and relatively selective. Differences in regeneration

potential (coppicing and seedling) in areas logged in 1999 and 1995 could have been

due to moisture availability and time since logging. Regeneration of B. plurijuga is

mainly by basal sprouting and less through propagation by seedling.

57

REFERENCES

Abbort, J.I.O and Homewood, K. 1999. A history of change: causes of miombo woodland decline in a protected area in Malawi. Journal of Applied Ecology 36, 422-433. Ashton, P.M.S., Samarasinghe, S.J., Gunatilleke, I.A.U.N and Gunatilleke, C.V.S. 1997. Role of legumes in release of successionally arrested grasslands in the central hills of Sri Lanka. Restoration Ecology 5, 36-43. Babaasa, D., Ellu, G., Kasangaki, A., Bitariho, R, and McNellage, A. 2004. Gap characteristics and regeneration in Bwindi Impenetrable National Park. African Journal of Ecology 42, 1365-2028. Baillie, I and Ashton, P.S. 1983. Soils and nutrients in Sarawak Forest. In: Sutton, S.L., Whitmore,T.C and Chadwick, A.C (eds.). Tropical rainforest: ecology and management Special publications series of the British Ecological Society Number 2, Blackwell Scientific Publications, London. Begon, M., Harper, J.L. and Townsend, C.R. 1996. Ecology. Blackwell, Oxford. Bell, R.H.V. 1982. The effects of soil nutrient availability on community structure in African ecosystems. In: Huntley, B.J. and Walker, B.H. (eds.). Ecology of tropical Savannas. Ecological studies 42, pp 193-216. Springer-Verlag, New York. Blackmore, A.C., Mentis, M.T., and Scholes, R.J. 1990. The origin and extent of nutrient-enriched patches within a nutrient-poor Savanna. Journal of Biogeography 17, 463-470. Bond and Van Wilgen, 1996. Fire and plants. Chapman and hall, London. Botha, J., Witkowski, E.T.F, and Shackleton, C.M.A.2000. A comparison of anthropogenic and elephant disturbance on Acacia xanthophloea populations in the lowveld. In: Shackleton, C.m. 2000. Stump size and the number of coppice shoots for selected Savanna tree species. South African Journal of Botany 66, 124-127. Bormann, F.H and Likens, G.E. 1979. Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York. Bradley, P.N and Dewees, P. 1993. Indigenous woodlands, agricultural production and household economy in communal areas. In: Bradley, P.N and McNamara (eds.). Living with Trees “Policies for Forestry Management in Zimbabwe”. World Bank Technical Paper № 210.

58

Brokaw, N.V.L. 1985a. Tree falls, regrowth and community structure in tropical forests. In: Pickett, S.T.A and White, P.S. (eds.). The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, New York. Campbell, D.G. 1988. Quantitative inventory of tropical forests. In: Campbell, D.G and Hammond, H.D (eds.). Floristic inventory of Tropical Countries. New York Botanical Garden, New York. Cannon, C.H., Peart, D.R., Leighton, M. and Kartawinata, K. 1994. The structure of lowland rainforest forest after selective logging in west Kalimantan. Forest Ecology Management 67, 49-68. Calvert, G.M. 1986a. The ecology and management of the Kalahari sand forest vegetation of south-western Zimbabwe. In Pierce, G.D. (ed.). The Zambezi Teak Forests. Proceedings of the First International Conference on the Teak Forests of Southern Africa, Livingstone, Zambia, 18-24 March. pp. 121-157. Chazdon, R.L. 1998. Tropical forests-Log “em or leave”em? Science 281, 1295-1296. Chapman, C. A and Chapman, L.J.1997. Forest regeneration in logged and unlogged areas of Kibale National Park. Biotropica 29, 396-412. Chenje, M., Sola, L, and Paleczny. D. (eds.). 1998. The state of Zimbabwe’s environment. Government of the Republic of Zimbabwe, Ministry of Mines, Environment and Tourism, Harare. Chidumayo, E.N. 1988a. Estimating fuelwood production and yield in regrowth dry miombo woodland in Zambia. Forest Ecology Management 24, 59-66. Chidumayo, E.N. 1994. Phenology and nutrition of miombo woodland trees in Zambia. Trees 9, 67-69. Chigwerewe, S.K. 1996. The ecology and management potentials of naturally regenerated Teak forests of the Kalahari-Namib sandstone formations in Zimbabwe. In: Mushove, P. T, Shumba, E.M, and Matose, F.(eds.). Sustainable Management of Indigenous Forests in the dry tropics. Proceedings of an International Conference, Kadoma, Zimbabwe, May 28–June. pp148 – 153. Forestry Commission, Harare. Childes, S.L, and Walker, B.H. 1987. Ecology and dynamics of the woody vegetation on the Kalahari Sands in Hwange National Park.Vegetation 72, 111-128. Chisumpa, S.M. 1984. Mukusi ecological associations and environmental effects. In: Piearce, G.D. (ed.). The Zambezi Teak Forests. Proceedings of the first International Conference on the Teak Forests of Southern Africa, Livingstone. Connell, J.H. and Slatyer, R.O. 1977. Mechanisms of succession in natural communities and their role in community stability and organisation. Am. Nat. 111, 1119-1144.

59

Coughenour, M.B and Ellis, J.E, 1993. Landscape and climatic control of woody vegetation in a dry tropical ecosystem: Turkana District, Kenya. Journal of Biogeography 20, 383-398. DeMeo, T. 1984.The case of natural regeneration system in managing the Zambezi teak forests. In: The Zambezi Teak Forests. Proceedings of the first International Conference on the Teak Forests of Southern Africa, Livingstone, 392-398. Denslow, J. S.1987. Tropical rain forest gaps and tree species diversity. Ann Rev. Ecology Syst. 18, 431-451. De Steven, R. 1991. Experiments on mechanisms of tree establishment in old-field succession: seedling emergence. Ecology 72, 1066-1075. Dye, P.J and Spear, P.T. The effect of bush clearing and rainfall variability on grass yield and composition in southwest Zimbabwe. Zimbabwe Journal of Agricultural Research 20, 103-117. Escudero, A., Garrido, M.V., Matias, M.D and Del Arco, D. 1988. Litter decomposition and nutrient dynamics in different environments in an open woodland. In: Verhoeven, J.T.A, Heil, G.W and Wegner, M.J.A (eds.). Vegetation structure in relation to carbon and nutrient cycling economy. SPB Academic Publishing, The Hague. Figueroa-Rangel, B.L and Olivera-Vargas, M. 2000. Regeneration patterns in relation to canopy species composition and site variables in mixed oak forests in the Sierra de Manantian Biosphere Reserve, Mexico. Ecological Research 15, 249-261. Gambiza, J. 2001. The Regeneration of Zambezi Teak after logging: Influence of Fire and Herbivory. Unpublished Ph D Thesis, University of Zimbabwe, Harare. Gauch, H.G. Jr. 1982. Multivariate analysis in community ecology. Cambridge Press, Cambridge. Gondo,P.C.1993. Fire management in the dermacated indigenous forests of Zimbabwe. In: Piearce,G.D and Gumbo, D.J (eds.). The ecology and management of indigenous forests in Southern Africa. Proceedings of an international symposium, Victoria Falls, Zimbabwe, 25-29 July 1992. Zimbabwe Forestry Commission, Harare. pp 336-371. Gotosa, T. N. 2002. Patterns of vegetation development on a 30 year abandoned sugar plantation in the Zambezi Valley. Unpublished M.Sc. Thesis, University of Zimbabwe, Harare.

60

Gutierrez, Armesto, J.J, and Aravena, J.C. 2004. Disturbance and regeneration dynamics of an old-growth North Patagonian rain forest in Chile Island, Chile. Journal of Ecology 92, 598-694. Guy, P.R. 1981b. Changes in the biomass and productivity of woodlands in the Sengwa Wildlife Research Area. Journal of Applied Ecology 18, 507-519. Hogberg, P. 1986. Rooting habits and mycorrhizas of Baikiaea plurijuga. In: Piearce, G.D. (ed.). The Zambezi Teak Forests. Proceedings of the first International Conference on the Teak Forests of Southern Africa, Livingstone, pp 92-100. Howlett, P.M.B and Davidson, D. 1996. Selective logging effects on abundance, diversity and composition of forests. For. Ecol. Manag. 23, 132-146. Huckabay, J.D. 1984. The Geography of Zambezi Teak. In: Piearce, G.D. (ed.). The Zambezi Teak Forests. Proceedings of the first International Conference on the Teak Forests of Southern Africa, Livingstone, pp 5-22. Huntly, T.1991. Herbivores and dynamics of communities and ecosystems. Ann. Rev. Ecol. Syste. 22, 477-503. Huston, M. 1979. A general hypothesis of species diversity. Amm. Nat. 113, 81-101. Judge,D.G 1986. The teak forests of Zimbabwe. In: Piearce, G.D. (ed.). The ecology and management of indigenous forests in Southern Africa, pp. 366-371. Proceedings of an International Symposium, Victoria Falls, 27-29 July 1992. Forestry Commission, Harare. Kasanene, J.M. and Murphy, P.G. 1987. Post-logging tree mortality and major branch losses in Kibale forest, Uganda. Forest Ecology Management 46, 295-307. Kikula, I.S. 1986. The influence of fire on composition on miombo woodland of SW Tanzania. Oikos 46, 317-324. Knowles, P and Grant, M.C. 1983. Age and size structure analyses of Engelman spruce, Ponderosa pine, Lodgepolepine and Limberpine in Colorado. Ecology 64, 1-9. Kobayashi,T., Hori, Y and Nomoto, S. 1997. Effects of Trampling and Vegetation removal on species diversity and macro-environment under different shade conditions. Journal of Vegetation Science 8, 873-880. Krebs, 1972. Ecology: The experimental analysis of distribution and abundance. Harper and Raw, New York. Liberman, D and Mingguang, L. 1992. Seedling recruitment patterns in a tropical dry forest in Ghana. Journal of Vegetation Science 3, 375-382.

61

Lieth, H. and Lohmann, M. 1993. Restoration of Tropical Ecosystems. Kluwer Academic Publishers, London. Lubke, R.A and Thatcher, F.M. 1983. Short term changes in the woody vegetation of Nylsvley. South African Journal of Botany 2, 85-97. Ludwig, J. A. and Reynolds, J.F. 1988. Statistical Ecology: A primer on methods and computing. John Wiley and Sons, New York. Magurran, A.E 1988. Ecological diversity and its measurements. Chapman and Hall, London. Makuku, S.J. 1990. Indigenous timber woodlots resuscitate natural ecology, provide more socio-economic benefits and are easier to establish and cheaper to maintain than exotic timber woodlots. Zimbabwe Science News 24, 91-93. Mapaure, I. 2001. The Influence of Elephants and Fire on the structure and Dynamics of Miombo Woodland in Sengwa Wildlife Research Area, Zimbabwe, Unpublished PhD Thesis, University of Zimbabwe, Harare. McGarigal, K., Cushman, S and Stafford. 2000. Multivariate Statistics for Wildlife and Ecology Research. Springer-Verlag, New York. Monela, G.C., O’kting’Ati, A and Kiwele, P.M. 1993. Socio-economic aspects of charcoal consumption and environmental consequences along the Da es Salaam-Morogoro highway. Tanzania. Forest Ecological Management 58, 249-258. Moukam, A and Nyakanou, D. 1997. The fertility status of surface soils in the humid forest zone of Southern Cameroon. African Crop Science Journal 5, 273-284. Mululuma, A.C.T. 1984. Possible alternate sites for Baikiaea plurijuga regeneration. In: Piearce, G.D. (ed.). The Zambezi Teak Forests. Proceedings of the first International Conference on the Teak Forests of Southern Africa, Livingstone, Zambia. Mushove, P.T, 1991. Indigenous hardwood based timber industry in Zimbabwe status. Unpublished paper presented to a seminar at the Forest Research Centre, Harare, 13 December 1991. Mushove, P.T., Gondo, P.C and Gumbie. 1993. The Effect of Silviricultural Cleaning on Diameter Increment in B. plurijuga. In: Piearce, G.D and Gumbo, D.J. (eds.). The Ecology and Management of Indigenous Forests in Southern Africa. Proceedings of an International Symposium, Victoria Falls, Zimbabwe, 27-29 July 1992, Forestry Commission, Harare pp 182-190. Nemarundwe, N, and Mbedzi, D. 1999. Social Development Studies for the Gwayi- Mbembesi settlement option study. Forestry Commission, Harare.

62

Ng, F.S.P. 1978. Strategies of establishment in Malayan forest trees. In: Tomlison, T.B and Zimmerman, M. H (eds.). Tropical trees as living systems. Cambridge University Press, Cambridge. Nyamapfene, K. 1991 Soils of Zimbabwe. Nehanda Press, Harare. Oberbauer, S.F and Strain, B.R. 1985. Effects of light regime on growth and physiology of Pentaclethra macroloba in Costa Rica. Journal of Ecology 1, 303-320. Osmaston, H.A. 1959. Working plan for the Kibale and itwera Forests. First revision. Period 1959-1965. Uganda Forest Department, Entebbe. Paciorek, C. J., Condit, R., Hubbell, S.P and Foster, R.B. 2000. The Demographics of resprouting in tree and shrub species of moist tropical forest. Forest Journal of Ecology 88, 765-777. Packham, J.R and Harding, D.T.L. 1982. Ecology of woodland processes. Edward Arnold, London. Palmer, M. J. 1993. Putting things together in even better order: the advantages of canonical correspondence analysis. Ecology 78, 2251-2230. Pellew, R.A. 1983. The impacts of elephant, giraffe and fire upon the Acacia tortilis woodlands of the Serengeti. African Journal of Ecology 21, 41-74. Piearce, G.D. 1986. How to save the Zambezi teak forests. Unasylva 152, 29-36. Piearce, G.D.1993. Natural Regeneration of Indigenous Trees: The key to their Successful Management. In: Piearce, G.D and Gumbo, D.J. (eds.). The Ecology and Management of Indigenous Forests in Southern Africa. Proceedings of an International Symposium, Victoria Falls, Zimbabwe, 27-29 July 1992, Forestry Commission, Harare pp 109-123. Pinard, M., Howlett, B. and Davidson, D.1996. Site conditions limit pioneer tree recruitment after logging of a dipterocarp forest in Sabah, Malaysia. Biotropica 28, 212-236. Rushworth, J.E. 1975. The floristic, physiognomic and biomass structure of Kalahari sand vegetation in relation to fire and frost in Wankie National Park, Rhodesia. Unpublished M.Sc. thesis, University of Rhodesia. Sabogal, C. 1992. Regeneration of tropical forests in Central America, with examples from Nicaragua. Journal of Vegetation Science 3, 407-416.

63

Scholes, R.J. and Parsons, D.A.B. 1997. The Kalahari transect: Research on Global change and sustainable developed in Southern Africa. IGBP Report 42. Stockholm, Sweden. Schwilk, D.W., Keeley, J.E and Bond, W.J. 1997. The Intermediate disturbance hypothesis does not explain fire and diversity patterns in fynbos. Plant Ecology 132, 77-84. Shackleton, C.M. 1994. Demograph and dynamics of the dominant woody species in a communal and protected area of the Eastern Transvaal lowveld. South African Journal of Botany 59, 569-574. Shackleton, C.M. 1997. The prediction of woody productivity in the Savanna biome. Unpublished PhD thesis. University of Witwatersrand, South Africa. Silvertown, J. 1987. Introduction to plant population ecology. Longman Scientific and Technical, New York. Stromgaard, P. 1992. Immediate and long-term effects of fire and ash-fertilisation on a Zambian miombo woodland soil. Agric. Ecosyst. Environ. 41, 19-37. Struhsaker, T.T., Lwanga, J. S. and Kasenene, J.M.1996. Elephants, selective logging and forest regeneration in the Kibale Forest, Uganda. Journal of Tropical Ecology 12, 45-64. Swaine, M.D and Liberman, D. 1987. The dynamics of tree populations in tropical forest. Journal of Tropical Ecology (Special issue) 289-375. ter Braak, C.J.F. 1986. Canonical correspondence analysis: A new eigenvector technique for multivariate direct gradient analysis. Ecology 67, 1167-1179. ter Braak, C. J. F. 1987. The analysis of vegetation-environment relationships by canonical correspondence analysis. Vegetation 69, 69-77. ter Braak, C. J. F and Smilauer. 1988. CANOCO Reference Manual and User’ s Guide for CANOCO for Windows. Software for Canonical Community Ordination. Ter Braak, C.J.F. 1995. Ordination. In: Jongman, R.H.G., ter Braak, C.J.F and Tongeren, O.F.R (eds.). Data analysis in community and landscape ecology. Cambridge University Press, Cambridge, pp 91-173. The Forest Survey in The Gwayi and Bembesi areas in the Republic of Zimbabwe Final Report, March 2001. Japan Forest Technical Association Kogyo Co Ltd, Forestry Commission, Harare. Theilade, I., Sekeli, P.M., Hald, S and Graudal, L. 2001. Conservation plan for genetic sources of Zambezi Teak (B. plurijuga) in Zambia, Humllback, Lusaka.

64

Timberlake, J.R., Nobanda, N and Mapaure,I. 1993. Vegetation Survey of Communal lands north and west of Zimbabwe. Kirkia 14, 171-270. Trollope, W.S.W. 1980. Controlling bush encroachment with fire in the Savanna area of South Africa. Proceedings of the Grassland Society of Southern Africa 15, 173-177. Tschaplinski, and Blake, T.J. 1989. Photosynthetic re-invigoration of leaves following shoot decapitation and accelerated growth of coppice shoots. Physiological Plant 75, 157-165. Guy, P.R. 1981. Changes in the biomass and productivity of woodlands in the Sengwa Wildlife Research Area. Journal of Applied Ecology 18, 507-519. Uhl, C., Clark, K., Dezzao, N and Maquino, P. 1988. Vegetation dynamics in Amazon tree fall gaps. Ecology 69, 751-763. Walker, B.H.1976. An approach to the monitoring changes in the composition and utilization of woodland and Savanna vegetation. South African Journal of Wildlife Research 6, 1-32. Walker, B.H., Stone, L., Henderson, L and Vernede, M. 1986. Size structure analysis of the dominant trees in a South African Savanna. South African Journal of Botany 52, 397-402.

Wilson, J.B. 1994. The ‘Intermediate Disturbance Hypothesis of species coexistence is based on patch dynamics. New Zealand Journal of Ecology 18, 176-181. Werner, B.H and Caswell, H. 1977. Population growth rates and age-vs-stage distribution models for teasel. Ecology 58, 1103-1111. Wolft, A and Debussche, M. 1999. Ants as seed dispersers in a Mediterranean Old Field Succession. Oikos 84, 443-452. http://www.ems.org/wildfires. Ecology of wildfires. 2001.Environmental Media Services, Washington DC. http://www.findarticles.com/p/articles/mi. Parker, G.R. 2003. Forest dynamics and Disturbance regimes: Studies from Temperate Evergreen- Deciduous Forests.

65

APPENDICES

Appendix 1: One-Way Analysis of Variance for variations in pH in Gwayi Forest Source DF SS MS F P pH 17 24.711 1.454 9.20 0.000 Error 27 4.267 0.158 Total 44 28.978 APPENDIX 2: One-Way Analysis of Variance for variation of Phosphorus in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest Source DF SS MS F P P 7 2.440 0.349 0.47 0.850 Error 37 27.471 0.742 Total 44 29.911 Appendix 3: One–Way Analysis of Variance for variation of Potassium in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest Source DF SS MS F P K 7 4.135 0.591 0.85 0.556 Error 37 25.776 0.697 Total 44 29.911 Appendix 4: One-Way Analysis of Variance for Nitrogen variation in an unlogged area, area logged in 1999 and area logged in 1995 Gwayi Forest Source DF SS MS F P N 11 15.128 1.375 3.07 0.006 Error 33 14.783 0.448 Total 44 29.911 Appendix 5: One–Way Analysis of Variance for diversity in an unlogged area, area logged in 1999 and area logged in 1995 Source DF SS MS F P Diversit 43 29.500 0.686 1.37 0.602 Error 1 0.500 0.500

66

Total 44 30.000 Appendix 6: Total area sampled and number of B. plurijuga in unlogged area, area logged in 1995 and area logged in 1999 in Gwayi Forest

Area Total area sampled Total number of B. plurijuga trees

Unlogged 7644 243

Logged in 1999 9947.5 239

Logged in 1995 9915 244

Appendix 7: Kruskal-Wallis Test results for Area versus Tree Density in an unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest Tree Den N Median Ave Rank Z 0.518186 2 2.500 30.5 0.83 0.556803 1 3.000 38.0 1.15 0.633614 1 2.000 23.0 0.00 0.708346 1 2.000 23.0 0.00 0.709265 1 1.000 8.0 -1.15 0.754232 1 3.000 38.0 1.15 0.926309 1 1.000 8.0 -1.15 0.932112 1 1.000 8.0 -1.15 0.952266 1 2.000 23.0 0.00 0.984980 1 2.000 23.0 0.00 1.019496 1 3.000 38.0 1.15 1.025915 1 3.000 38.0 1.15 1.044186 1 3.000 38.0 1.15 1.085086 1 2.000 23.0 0.00 1.105005 1 3.000 38.0 1.15 1.110382 1 1.000 8.0 -1.15 1.133170 1 1.000 8.0 -1.15 1.154105 1 1.000 8.0 -1.15 1.165803 1 1.000 8.0 -1.15 1.195611 1 2.000 23.0 0.00 1.201209 1 1.000 8.0 -1.15 1.202353 1 2.000 23.0 0.00 1.221131 1 2.000 23.0 0.00 1.236533 1 1.000 8.0 -1.15 1.248825 1 2.000 23.0 0.00 1.248827 1 2.000 23.0 0.00 1.276066 1 3.000 38.0 1.15 1.344007 1 1.000 8.0 -1.15 1.369875 1 3.000 38.0 1.15 1.369995 1 3.000 38.0 1.15 1.369995 1 3.000 38.0 1.15 1.376622 1 2.000 23.0 0.00 1.395896 1 2.000 23.0 0.00 1.402567 1 1.000 8.0 -1.15 1.431954 1 3.000 38.0 1.15 1.441193 1 1.000 8.0 -1.15

67

1.444059 1 3.000 38.0 1.15 1.456586 1 2.000 23.0 0.00 1.506913 1 1.000 8.0 -1.15 1.519820 1 3.000 38.0 1.15 1.566603 1 2.000 23.0 0.00 1.593382 1 3.000 38.0 1.15 1.811675 1 1.000 8.0 -1.15 1.9736 1 1.000 8.0 -1.15 Overall 45 23.0 H = 38.48 DF = 43 P = 0.668 H = 43.27 DF = 43 P = 0.460 (adjusted for ties) Appendix 8: Kruskal-Wallis Test results for stump density versus logged area Area N Median Ave Rank Z Logged in 1995 15 0.005180 16.2 0.44 Logged in 1999 15 0.003900 14.8 -0.44 Overall 30 15.5 H = 0.19 DF = 1 P = 0.663 Df= 20 Appendix 9: One-way Analysis of Variance for species Richness Source DF SS MS F P Time 2 1.91 0.96 0.23 0.796 Error 42 174.67 4.16 Total 44 176.58 Appendix 10: One-way Analysis of Variance for comparing basal area in the unlogged area, area logged in 1999 and area logged in 1995 Source DF SS MS F P Time 2 0.1646 0.0823 2.38 0.105 Error 42 1.4544 0.0346 Total 44 1.6190 Appendix 11: Number of stumps per plot in an area logged in 1999 and area logged in 1995

Plot Stumps/N

Stumps/C

16 5 1 17 0 0 18 1 2 19 2 1 20 4 1

68

21 4 0 22 2 0 23 1 1 24 2 1 25 3 0 26 3 1 27 1 1 28 3 1 29 3 0 30 5 1 31 1 0 32 2 1 33 1 3 34 1 1 35 2 1 36 1 3 37 1 1 38 1 2 39 2 1 40 2 2 41 2 5 42 1 1 43 2 1 44 1 2 45 0 3

Key Stumps/N-Stumps not regenerating Stumps/C-Stumps regenerating Appendix 12: Chi-Square Test for comparing stump regeneration in the area logged in 1999 and in the area logged in 1995 area 1995 1999 Total 1 11 27 38 Reg. (19.59) (18.41) Status 2 39 20 59 (30.41) (28.59) Total 50 47 97 Chi-Sq = 3.765 + 4.005 + 2.425 + 2.580 = 12.775 DF = 1, P-Value = 0.000 Key Reg. status- Regeneration status

69

(..)-E value Site 1- Area logged in 1995 Site 2-Area logged in 1999 Appendix 13: Tree density, richness and diversity from sampling plots in Gwayi Forest

Plot number

Tree density H

Richness

1 0.02 1.154105 7 2 0.03 1.165804 7 3 0.0209 1.506913 7 4 0.0273 1.811675 11 5 0.0197 1.9736 11 6 0.03478 1.402567 8 7 0.0545 0.709265 5 8 0.0588 0.932112 4 9 0.05313 1.344007 5 10 0.048 1.441194 9 11 0.035 1.20121 6 12 0.03214 1.236534 6 13 0.02833 1.110382 5 14 0.0475 0.926309 5 15 0.0459 1.13317 5 16 0.02957 1.376622 8 17 0.025 1.248827 7 18 0.0237 0.708347 4 19 0.02267 1.202354 6 20 0.0207 1.195612 6 21 0.03 0.984981 4

22 0.02727 1.395896 8

23 0.0323 1.248826 7 24 0.0133 1.456586 8 25 0.02326 1.221131 8 26 0.0283 1.085086 5 27 0.0536 0.633614 4

28 0.03448 0.952266 6

29 0.0298 0.518186 3 30 0.015075 1.566604 11 31 0.02454 1.105006 6 32 0.017 1.444059 7 33 0.029 1.369995 8 34 0.02677 1.593382 10 35 0.03 1.276066 8 36 0.0278 0.556804 8 37 0.027778 1.369875 8 38 0.019 1.044187 6 39 0.02459 1.431954 7 40 0.02133 1.519821 8 41 0.0411 1.369996 8 42 0.029 1.025916 5 43 0.01887 1.019497 5 44 0.01935 0.754232 5

70

45 0.022368 0.518186 3 APPENDIX 14: Measurements of environmental variables hypothesized to be influencing population structure and regeneration of B. plurijuga in Gwayi Forest Plot pH

(0.01MCaCl2)

Ppm P Ppm N Me% K

1 5.6 3 7 0.09 2 5.2 3 4 0.08 3 5.3 3 5 0.07 4 5.3 3 10 0.10 5 5.3 1 9 0.08 6 5.7 8 5 0.08 7 6.2 5 14 0.09 8 5.6 3 8 0.08 9 6.1 13 8 0.11 10 5.6 5 5 0.12 11 5.9 8 7 0.09 12 5.8 8 4 0.11 13 5.5 3 6 0.11 14 5.5 3 7 0.09 15 5.9 8 3 0.08 16 5.8 3 4 0.11 17 5.8 13 3 0.09 18 6.1 18 2 0.10 19 5.9 5 6 0.10 20 5.8 5 0 0.12 21 6.1 23 7 0.10 22 6.0 8 3 0.10 23 5.8 23 2 0.08 24 6.0 5 5 0.08 25 5.9 5 1 0.09 26 6.0 5 2 0.17 27 6.0 5 1 0.27 28 6.0 5 0 0.10 29 6.5 10 3 0.08 30 5.9 5 3 0.10 31 6.4 3 2 0.12 32 6.2 3 1 0.08 33 6.7 1 1 0.10 34 7.0 1 0 0.07 35 6.8 3 1 0.12 36 7.3 3 1 0.11 37 7.1 1 5 0.08 38 7.0 1 0 0.10 39 7.0 3 3 0.08 40 7.0 3 0 0.10 41 7.0 3 0 0.09 42 6.9 13 0 0.17 43 7.1 5 0 0.10 44 7.0 8 4 0.10 45 6.8 3 8 0.09

71

Appendix 15: B. plurijuga stem density, basal area and stump density calculated from sampling plots in Gwayi Forest Plot number

Tree density/m2

Basal area/m2

Stump density/m2

1 0.02 0.584 0 2 0.03 0.258 0 3 0.0209 0.513 0 4 0.0273 0.224 0 5 0.0197 0.072 0 6 0.03478 0.862 0 7 0.0545 0.523 0 8 0.0588 0.194 0 9 0.05313 0.84 0 10 0.048 0.327 0 11 0.035 0.515 0 12 0.03214 0.485 0 13 0.02833 0.425 0 14 0.0475 0.676 0 15 0.0459 0.231 0 16 0.02957 0.439 0.001739 17 0.025 0.46 0 18 0.0237 0.491 0.001481 19 0.02267 0.438 0.004 20 0.0207 0.738 0.00518 21 0.03 0.413 0.008 22 0.02727 0.495 0.03636 23 0.0323 0.112 0.0043 24 0.0133 0.568 0.0025 25 0.02326 0.266 0.00465 26 0.0283 0.2383 0.0053 27 0.0536 0.142 0.0071 28 0.03448 0.135 0.0069 29 0.0298 0.2194 0.00526 30 0.015075 0.3084 0.00603 31 0.02454 0.408 0.001444 32 0.017 0.275 0.003 33 0.029 0.258 0.007767 34 0.02677 0.467 0.00315 35 0.03 0.248 0.005 36 0.0278 0.319 0.00696 37 0.027778 0.3324 0.0037 38 0.019 0.286 0.00382 39 0.02459 0.173 0.0049 40 0.02133 0.618 0.005333 41 0.0411 0.237 0.019178 42 0.029 0.388 0.005455 43 0.01887 0.143 0.00377 44 0.01935 0.116 0.00387

72

45 0.022368 0.246 0.0039 Appendix 16: Chi square test for comparing basal size classes in unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest area 1 2 3 Total 1 163 156 137 456 159.14 143.07 153.79 2 93 69 95 257 89.69 80.63 86.67 3 25 29 48 102 35.60 32.00 34.40 4 14 11 7 32 11.17 10.04 10.79 5 1 2 0 3 1.05 0.94 1.01 8 1 0 0 1 0.35 0.31 0.34 Total 297 267 287 851 Chi-Sq = 0.093 + 1.169 + 1.832 + 0.122 + 1.678 + 0.800 + 3.155 + 0.282 + 5.377 + 0.718 + 0.092 + 1.332 + 0.002 + 1.191 + 1.012 + 1.214 + 0.314 + 0.337 = 20.721 DF = 10 Appendix 17: Geographical positions of sampling plots in the unlogged area, area logged in 1999 and area logged in 1995 in Gwayi Forest Plot number UTM Position 1 586501, 7890520 2 589278, 7894532 3 590124, 7893097 4 589278, 7892532 5 591179, 7893097 6 588010, 7892532 7 590124, 7892532 8 587071, 7891183 9 589279, 7893097

73

10 588010, 7891183 11 587508, 7890520 12 586501, 7891183 13 588010, 7893097 14 587071, 7891532 15 589278, 7891183 16 591179, 7886202 17 593926, 7886202 18 594155, 78887267 19 596439, 7888963 20 598071, 7887267 21 597071, 7887267 22 595060, 7886962 23 594155, 7885840 24 594155, 7884836 25 596439, 7884836 26 599050, 7885840 27 599050, 7886962 28 597071, 7886962 29 596439, 7885840 30 600497, 7886962 31 580524, 7899321 32 581267, 7899356 33 583437, 7898356 34 584258, 7899321 35 582739, 7899321 36 580345, 7897321 37 580524, 7897321 38 580524, 7896438 39 582739, 7897321 40 585536, 7897321 41 584258, 7896438 42 582739, 7896321 43 581267, 7895594 44 582739, 789594 45 586501, 7897321

Appendix 18: CCA results for comparing species-environmental relations SPEC AX1 1.0000 SPEC AX2 -.0239 1.0000 SPEC AX3 .0481 -.0620 1.0000 SPEC AX4 .1179 -.1017 .0743 1.0000 ENVI AX1 .8895 .0000 .0000 .0000 1.0000 ENVI AX2 .0000 .7613 .0000 .0000 .0000 1.0000 ENVI AX3 .0000 .0000 .7526 .0000 .0000 .0000 1.0000 ENVI AX4 .0000 .0000 .0000 .6705 .0000 .0000 .0000 1.0000 pH -.7992 .2159 .2520 .0070 -.8985 .2837 .3349 .0105 P .4154 .6520 .1640 -.0199 .4670 .8565 .2180 -.0296 N .5416 -.3802 .4461 -.1132 .6088 -.4995 .5928 -.1688 K -.0209 .1117 .0186 .6628 -.0235 .1467 .0247 .9886

74

SPEC AX1 SPEC AX2 SPEC AX3 SPEC AX4 ENVI AX1 ENVI AX2 ENVI AX3 ENVI AX4 Axes 1 2 3 4 Total inertia Eigenvalues : .232 .128 .100 .067 3.798 Species-environment correlations : .890 .761 .753 .670 Cumulative percentage variance of species data : 6.1 9.5 12.1 13.9 of species-environment relation: 44.0 68.4 87.3 100.0 Sum of all unconstrained eigenvalues 3.798 Sum of all canonical eigenvalues .526 Appendix 19: Summary of DCA results Axes 1 2 3 4 Total inertia Eigenvalues : .340 .248 .174 .132 3.798 Lengths of gradient : 7.331 5.418 5.884 5.963 Cumulative percentage variance of species data : 8.9 15.5 20.1 23.5 Sum of all unconstrained eigenvalues 3.798 Appendix 20: Summary of Permutation test results Variable F-value p pH 2.50 0.0050 P 1.72 0.0250 N 1.33 0.1350 K 0.84 0.6100 Appendix 21: One-way ANOVA for comparing seedlings in the sampling plots Source DF SS MS F P Plot 2 2.80 1.40 0.84 0.439 Error 42 70.00 1.67 Total 44 72.80

75

Appendix 22: One-way ANOVA for comparing saplings in the sampling plots Source DF SS MS F P Plot 2 1.24 0.62 0.47 0.627 Error 42 55.33 1.32 Total 44 56.58 Appendix 23: Abundance of woody species in an Unlogged area, area logged in 1995 and area logged area logged in 1999 Species Unlogged Logged in 1995 Loggedin 1999 Diplorhynchus condylocarpon 10 3 28 Acacia species 5 19 1 Albizia antunesiana 1 0 0 Albizia tanganyincensis 0 1 0 Grewia species 43 18 30 Baikiaea plurijuga 299 285 268 Burkea Africana 2 2 11 Baphia massieensis 0 0 0 Bauhinia petersiana 0 1 0 Croton gratissimus 41 26 0 Commiphora mollis 69 46 24 Commiphora pyranthoides 0 0 1 Combretum apiculatum 0 16 6 Combretum collinum 9 9 15 Combretum hereroense 0 1 0 Combretum molle 27 14 11 Combretum zeyheri 0 0 6 Combretum species 50 4 8 Vangueria apiculata 17 0 8 Indigofera species 4 0 0 Guibourtia coleosperma 0 3 11 Grewia retinervis 17 52 19 Dichapetulum rhodesiana 6 0 1 Hymenodictyon species 44 11 2 Commiphora malorthii 3 2 1 Ochna pulchra 0 4 19 Pseudolachnostylis maprouneifolia 5 3 5 Pterocarpus angolensis 9 5 5 Vitex pyos 35 0 0 Schinziophyton rautanenii 6 0 1 Strychnos pungens 2 1 44 Strychnos spinosa 0 3 2 Vangueria infausta 0 3 0 Ximenia Americana 1 7 0 Grewia flavescens 2 0 0 Rhus tenuinervis 0 0 0 Solanum species 17 0 3 Protorhus species 1 0 0 Grewia monticola 5 1 0

76

Strychnos madagascariensis 0 0 0 Ximenia caffra 0 0 1 Heteropogon melanocarpus 0 4 0 Afzelia quanzensis 0 0 0 Commiphora longibracteata 0 0 0 Jasminum stenobolum 0 1 0 Xylopia odoratissima 0 0 0 Terminalia sericea 47 4 40

Plot/Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 18 19 20 21 22 23 24 25 26 27 28 Diplorhynchus condylocarpon

1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0

Appendix 24: Presence/absence data used for CCA and DCA analysis

16 0 1

Acacia species 0 0 1 1 1 0 0 1 0 0 0 0 0 0 1 0 1 1 1 0 1 1 1 1 0 0 1 Albizia antunesiana 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Albizia tanganyicensis 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 Grewia species 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 Baikiaea plurijuga 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 Burkea africana 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 Baphia massiensis 1 0 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 0 1 0 0 0 0 0 0 Bauhinia petersiana 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Croton gratissimus 0 0 1 1 0 1 1 1 0 1 0 0 1 1 1 0 0 0 0 1 1 1 0 1 0 1 0 0 Commiphora mollis 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 1 0 1 1 1 1 1 Commiphora pyranthoides

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0

Combretum apiculatum 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 1 Combretum collinum 1 0 0 0 1 1 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 Combretum hereroense 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 Combretum molle 1 1 1 1 1 0 0 0 0 0 0 0 1 0 0 1 1 1 1 1 0 1 0 0 0 0 0 0 Combretum zeyheri 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Combretum species 1 1 0 0 1 0 0 1 0 0 0 1 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 Vitex pyos 0 1 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 Vangueria apiculata 0 0 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Dichapetulum rhodesiana

0 0 0 1 1 0 1 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 1 0 1 0 0

Guibourtia coleosperma 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 Grewia retinervis 0 0 0 1 0 1 0 1 0 0 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 1 1 0 Combretum species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 Hymenodictyon species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 Commiphora marlothii 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 Ochna pulchra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pseudolachnostylis maprouneifolia

0 0 0 1 1 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0

Pterocarpus angolensis 1 1 1 1 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 Schziphyton rautanenii 0 1 0 0 0 0 1 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Strychnos pungens 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 Strychnos spinosa 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 Terminalia sericea 1 1 1 1 1 1 0 1 0 1 0 1 1 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0

0

77

Vangueria infausta 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 0 0 Ximenia americana 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 1 1 Grewia flavescens 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Rhus tenuinervis 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 Solanum species 0 0 1 0 1 1 1 1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Protorhus species 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 Grewia monticola 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Strychnos madagagascariensis

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Ximenia caffra 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Heteropogon melanocarpus

0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Indigofera species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 Afzelia quanzensis 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Commiphora longibracteata

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Jasminum stenolobum 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Xylopia odoratissima 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0

Plot/species 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Diplorhynchus condylocarpon 0 0 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 Acacia species 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 Albizia antunesiana 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Albizia tanganyicensis 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Grewia species 1 0 1 1 1 1 0 0 1 0 1 1 1 0 1 1 1 Baikiaea plurijuga 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Burkea africana 0 1 0 0 1 0 0 0 1 0 1 0 1 0 0 1 1 Baphia massiensis 0 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 Bauhinia petersiana 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Croton gratissimus 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Commiphora mollis 0 1 0 0 0 0 0 0 1 1 1 1 1 1 1 0 1 Commiphora pyranthoides 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 Combretum apiculatum 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 Combretum collinum 1 1 0 0 1 0 1 1 1 1 1 0 1 0 1 0 0 Combretum hereroense 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Combretum molle 0 0 1 1 1 1 1 1 1 0 0 0 0 0 0 1 0 Combretum zeyheri 0 0 1 0 0 1 1 1 0 1 0 0 0 0 0 0 0 Combretum species 0 0 0 1 0 0 0 0 0 1 0 0 1 1 0 1 0 Vitex pyos 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Vangueria apiculata 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Dichapetulum rhodesiana 1 1 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 Guibourtia coleosperma 0 1 0 0 0 1 0 1 0 1 0 1 0 0 0 1 1 Grewia retinervis 1 1 1 1 0 1 1 1 0 1 1 1 0 1 0 0 1 Combretum species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Hymenodictyon species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Commiphora marlothii 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Ochna pulchra 0 0 1 0 1 0 0 1 1 1 0 1 1 1 1 0 0 Pseudolachnostylis maprouneifolia

1 0 0 0 1 0 0 0 0 0 1 0 0 0 1 0 1

78

Pterocarpus angolensis 1 0 0 0 0 0 1 0 0 1 1 0 1 0 0 0 0 Schziphyton rautanenii 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 Strychnos pungens 0 0 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 Strychnos spinosa 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 Terminalia sericea 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 Vangueria infausta 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Ximenia americana 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Grewia flavescens 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Rhus tenuinervis 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Solanum species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 Protorhus species 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 Grewia monticola 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Strychnos madagagascariensis 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Ximenia caffra 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 Heteropogon melanocarpus 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Indigofera species 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Afzelia quanzensis 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Commiphora longibracteata 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 Jasminum stenolobum 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Xylopia odoratissima 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0